arXiv subjects

Marko Čanađija

Publications and source records attributed to Marko Čanađija.

10 recordsLinked to original sources

arXiv2026

On Potentials and Complementary Potentials in One-Dimensional Nonlocal Integral Formulations

The present research presents potentials and complementary potentials used in the one-dimensional nonlocal integral formulations. The pure stress and the pure strain nonlocal formulations were considered. While the potential used in the strain driven formulation is well known, the complementary potential has not yet been presented in the literature. The same applies to the stress driven formulation. The equivalent formulations are obtained by resorting to the Legendre transformation, and their equivalence is proved. It is also shown that these results can be used to postulate a novel potential, i.e. a kind of mixed stress-strain potential, which is, however, as ill-conditioned as the pure strain-driven formulation. Finally, an example is given that practically confirms that the stress-driven formulations resulting from the potential and the complementary potential are equivalent.

physics.class-ph↗

arXiv2026

Modeling isotropic polyconvex hyperelasticity by neural networks -- sufficient and necessary criteria for compressible and incompressible materials

This work investigates different sufficient and necessary criteria for hyperelastic, isotropic polyconvex material models, focusing on neural network implementations for compressible and incompressible materials. Furthermore, the expressiveness, accuracy, simplicity as well as the efficiency of those models is analyzed. This also enables an assessment of the practical applicability of the models. Convex Signed Singular Value Neural Networks (CSSV-NNs) are applied to compressible materials and tailored to incompressibility (inc-CSSV-NNs), resulting in a universal approximation for frame-indifferent, isotropic polyconvex energies for the compressible as well as incompressible case. While other existing approaches also guarantee frame-indifference, isotropy and polyconvexity, they impose too restrictive constraints and thus limit the expressiveness of the model. This is further substantiated by numerical examples of several, well-established classical models (Neo-Hooke, Mooney-Rivlin, Gent and Arruda-Boyce) and Treloar's experimental data. Moreover, the numerical examples include an explicitly constructed energy function that cannot be approximated by neural networks constrained by Ball's criterion for polyconvexity. This substantiates that Ball's criterion, though sufficient, is not necessary for polyconvexity.

cs.CE↗

arXiv2026

Physics-Informed Neural Networks for Nonlocal Beam Eigenvalue Problems

The present study investigates the dynamics of nonlocal beams by establishing a consistent stress-driven integral elastic using the Physics-Informed Neural Network (PINN) approach. Specifically, a PINN is developed to compute the first eigenfunction and eigenvalue arising from the underlying sixth-order ordinary differential equation. The PINN is based on a feedforward neural network, with a loss function composed of terms from the differential equation, the normalization condition, and both boundary and constitutive boundary conditions. Relevant eigenvalues are treated as separate trainable variables. The results demonstrate that the proposed method is a powerful and robust tool for addressing the complexity of the problem. Once trained, the neural network is less computationally intensive than analytical methods. The obtained results are compared with benchmark analytical solutions and show strong agreement.

physics.class-ph↗

arXiv2024

Recovering Mullins damage hyperelastic behaviour with physics augmented neural networks

The aim of this work is to develop a neural network for modelling incompressible hyperelastic behaviour with isotropic damage, the so-called Mullins effect. This is obtained through the use of feed-forward neural networks with special attention to the architecture of the network in order to fulfil several physical restrictions such as objectivity, polyconvexity, non-negativity, material symmetry and thermodynamic consistency. The result is a compact neural network with few parameters that is able to reconstruct the hyperelastic behaviour with Mullinstype damage. The network is trained with artificially generated plane stress data and even correctly captures the full 3D behaviour with much more complex loading conditions. The energy and stress responses are correctly captured, as well as the evolution of the damage. The resulting neural network can be seamlessly implemented in widely used simulation software. Implementation details are provided and all numerical examples are performed in Abaqus.

physics.comp-ph↗

arXiv2024

Data-driven methods for computational mechanics: A fair comparison between neural networks based and model-free approaches

We present a comparison between two approaches to modelling hyperelastic material behaviour using data. The first approach is a novel approach based on Data-driven Computational Mechanics (DDCM) that completely bypasses the definition of a material model by using only data from simulations or real-life experiments to perform computations. The second is a neural network (NN) based approach, where a neural network is used as a constitutive model. It is trained on data to learn the underlying material behaviour and is implemented in the same way as conventional models. The DDCM approach has been extended to include strategies for recovering isotropic behaviour and local smoothing of data. These have proven to be critical in certain cases and increase accuracy in most cases. The NN approach contains certain elements to enforce principles such as material symmetry, thermodynamic consistency, and convexity. In order to provide a fair comparison between the approaches, they use the same data and solve the same numerical problems with a selection of problems highlighting the advantages and disadvantages of each approach. Both the DDCM and the NNs have shown acceptable performance. The DDCM performed better when applied to cases similar to those from which the data is gathered from, albeit at the expense of generality, whereas NN models were more advantageous when applied to wider range of applications.

cs.CE↗

arXiv2024

Carbon Nanotubes as a Basis of Metamaterials and Nanostructures: Crafting via Design Optimization

Nanotruss structures made of carbon nanotubes are investigated in two conceptual applications: either as building blocks of metamaterials or for nanostructural applications. The nanotrusses are optimized for different purposes, including various loadings, boundary conditions, parameterizations, objectives and constraints used to formulate optimization problems. The procedure relies on a recently developed framework consisting of molecular dynamics simulations, neural networks and finite elements. This framework is now used in the design optimization of nanostructures and the performances of different popular heuristic optimization methods are compared. Five applications of nanotrusses made of carbon nanotubes are analyzed in detail to investigate the mechanical behavior of such structures and the efficiency of the optimizations. Besides an introductory example, the design of an energy trapping carbon nanotube nanotruss, an auxetic nanotruss, a cantilever nanotruss and the maximization of the compressive strength of a metamaterial are presented. It is shown that the exceptional mechanical properties of carbon nanotubes can indeed be exploited for the development of structures and materials with extraordinary mechanical properties. Although hampered by material and geometrical nonlinearity of the problem, most of the tested optimization methods have proven to be a good choice for the design of such materials and structures.

cond-mat.mtrl-sci↗

arXiv2023

A computational framework for nanotrusses: input convex neural networks approach

The present research aims to provide a practical numerical tool for the mechanical analysis of nanoscale trusses with similar accuracy to molecular dynamics (MD). As a first step, MD simulations of uniaxial tensile and compression tests of all possible chiralities of single-walled carbon nanotubes up to 4 nm in diameter were performed using the AIREBO potential. The results represent a dataset consisting of stress/strain curves that were then used to develop a neural network that serves as a surrogate for a constitutive model for all nanotubes considered. The cornerstone of the new framework is a partially input convex integrable neural network. It turns out that convexity enables favorable convergence properties required for implementation in the classical nonlinear truss finite element available in Abaqus. This completes a molecular dynamics-machine learning-finite element framework suitable for the static analysis of large, nanoscale, truss-like structures. The performance is verified through a comprehensive set of examples that demonstrate ease of use, accuracy, and robustness.

physics.app-ph↗

arXiv2021

Dynamic behavior of nanobeams under axial loads: Integral elasticity modeling and size-dependent eigenfrequencies assessment

In this article, eigenfrequencies of nano-beams under axial loads are assessed by making recourse to the well-posed stress-driven nonlocal model (SDM) and strain-driven two-phase local/nonlocal formulation (NstrainG) of elasticity and Bernoulli-Euler kinematics. The developed nonlocal methodology is applicable to a wide variety of nano-engineered materials, such as carbon nanotubes, and modern small-scale beam-like devices of nanotechnological interest. Eigenfrequencies calculated using SDM, are compared with NstrainG and other pertinent results in literature obtained by other nonlocal strategies. Influence of nonlocal thermoelastic effects and initial axial force (tension and compression) on dynamic responses are analyzed and discussed. Model hardening size effects from stress-driven approach is compared to model softening size effects from strain-driven two-phase local/nonlocal approach for increasing nonlocal parameters.

physics.app-ph↗

arXiv2020

On thermomechanics of multilayered beams

In this paper, the mechanical behavior of multilayered small-scale beams in nonisothermal environment is investigated. Scale phenomena are modeled by means of the mathematically well-posed and experimentally consistent stress-driven integral formulation of elasticity. The present research extends the treatment in [1] confined to elastically homogeneous nano-scopic structures. It is shown that the non-locality leads to a complex coupling between axial and transverse elastic displacements. Such a size-dependent phenomenon makes the solution of the relevant nonlocal thermoelastostatic problem, governed by a system of two ordinary differential equations with ten standard boundary conditions and non-classical constitutive boundary conditions, significantly more involved with respect to treatments in literature. Thus, a novel solution methodology, based on Laplace transforms, is proposed and illustrated by examining simple structural schemes of current applicative interest in Nanomechanics and Nanotechnology.

physics.app-ph↗

arXiv2019

Nonlocal integral thermoelasticity: a thermodynamic framework for functionally graded beams

An effective nonlocal integral formulation for functionally graded Bernoulli-Euler beams in nonisothermal environment is developed. Both thermal and mechanical loadings are accounted for. The proposed model, of stress-driven integral type, is shown to be governed by a thermodynamically consistent differential problem with proper constitutive boundary conditions. The new thermoelastic strategy is illustrated by investigating a set of examples. It is demonstrated that in nonisothermal statically indeterminate problems rather complex structural behaviours can appear and that both the shift of the neutral surface and nonlocality have a dominating influence at small-scales.

physics.app-ph↗